American Journal of Infection Control xxx (2014) 1-4

Contents lists available at ScienceDirect

American Journal of Infection Control

American Journal of Infection Control

journal homepage: www.ajicjournal.org

Major article

The use of real-time feedback via wireless technology to improve hand hygiene compliance Alexandre R. Marra MD a, *, Thiago Zinsly Sampaio Camargo MD b, Thyago Pereira Magnus RN b, Rosangela Pereira Blaya RN b, Gilson Batista dos Santos RN b, Luciana Reis Guastelli RN b, Rodrigo Dias Rodrigues RN b, Marcelo Prado BE c, Elivane da Silva Victor PhD a, Humberto Bogossian MD b, Julio Cesar Martins Monte MD a, Oscar Fernando Pavão dos Santos MD d, Carlos Kazume Oyama BE e, Michael B. Edmond MD, MPH, MPA f a

Instituto Israelita de Ensino e Pesquisa Albert Einstein, Hospital Israelita Albert Einstein, São Paulo, Brazil Intensive Care Unit, Hospital Israelita Albert Einstein, São Paulo, Brazil c Division of Research and Development, I-HealthSys, São Carlos, Brazil d Division of Medical Practice, Hospital Israelita Albert Einstein, São Paulo, Brazil e Division of Logistics and Supplies, Hospital Israelita Albert Einstein, São Paulo, Brazil f Department of Internal Medicine, Virginia Commonwealth University School of Medicine, Richmond, VA b

Key Words: Feedback loop Electronic handwash counters Zigbee RFID Alcohol hand rub Quasiexperimental study Innovation Health care worker performance Positive deviance Communication system

Background: Hand hygiene (HH) is widely regarded as the most effective preventive measure for health care-associated infection. However, there is little robust evidence on the best interventions to improve HH compliance or whether a sustained increase in compliance can reduce rates of health care-associated infection. Methods: To evaluate the effectiveness of a real-time feedback to improve HH compliance in the inpatient setting, we used a quasiexperimental study comparing the effect of real-time feedback using wireless technology on compliance with HH. The study was conducted in two 20-bed step-down units at a private tertiary care hospital. Phase 1 was a 3-month baseline period in which HH counts were performed by electronic handwash counters. After a 1-month washout period, a 7-month intervention was performed in one step-down unit while the other unit served as a control. Results: HH, as measured by dispensing episodes, was significantly higher in the intervention unit (90.1 vs 73.1 dispensing episodes/patient-day, respectively, P ¼ .001). When the intervention unit was compared with itself before and after implementation of the wireless technology, there was also a significant increase in HH after implementation (74.5 vs 90.1 episodes/patient-day, respectively, P ¼ .01). There was also an increase in mean alcohol-based handrub consumption between the 2 phases (68.9 vs 103.1 mL/patient-day, respectively, P ¼ .04) in the intervention unit. Conclusion: We demonstrated an improvement in alcohol gel usage via implementation of real-time feedback via wireless technology. Copyright Ó 2014 by the Association for Professionals in Infection Control and Epidemiology, Inc. Published by Elsevier Inc. All rights reserved.

Hand hygiene (HH) is widely regarded as the most effective preventive measure for health care-associated infection (HAI).1

* Address correspondence to Alexandre R. Marra, MD, Avenue Albert Einstein, 627-Bloco A1-1o andar, Room 108, Morumbi, São Paulo, Brazil 05651-901. E-mail address: [email protected] (A.R. Marra). Supported by a grant from GOJO Latin America to implement the Zigbee wireless system in the intervention unit. Conflicts of interest: Marcelo Prado works for iHealthSys. The remaining authors report no conflicts.

Many strategies for improving HH compliance have been adopted in different studies1-4 with a variety of interventions.5-7 However, there is little robust evidence regarding the best interventions to improve HH compliance or to determine whether a sustained increase in compliance can reduce rates of HAI.1,3 Improving HH compliance is one of the performance improvement objectives of our institution, especially in units serving critically ill patients. In our intensive care unit, we have personnel who perform HH observations. Although direct observation has been considered the gold standard method, only a very small fraction of

0196-6553/$36.00 - Copyright Ó 2014 by the Association for Professionals in Infection Control and Epidemiology, Inc. Published by Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.ajic.2014.02.006

2

A.R. Marra et al. / American Journal of Infection Control xxx (2014) 1-4

HH opportunities can be observed (approximately 1% in our hospital).8 Observers can follow health care workers (HCWs) to perform direct HH observations; however, given that our stepdown rooms are private, HCWs would be prompted to clean their hands when observers were so close to them (Hawthorne effect), and thus it would not represent real-world conditions. Moreover, having observers walk into patient rooms violates patient privacy. HH episodes can be recorded by electronic handwash counters for alcohol gel.8-10 The electronic handwashing counter is an important tool for collecting information about HH, giving us the possibility to provide feedback to HCWs about their HH performance. However, feedback of product use has not resulted in significant improvement in HH.5 Other measures, including positive deviance7 for developing accountability among HCWs, should be considered to increase and sustain HH compliance. Feedback loops are profoundly effective tools for changing human behavior. This is based on a simple premise: give people information about their actions in near real time then show them how to transform those actions into better behaviors.11 The purpose of this study was to prospectively evaluate compliance with HH in 2 similar adult step-down units (SDU) using electronic handwash counters with the application of a feedback loop strategy using wireless technology. METHODS The study was performed in 2 medical-surgical SDU in a 610-bed, private, tertiary care hospital in São Paulo, Brazil. The units have the same physical layout, and each have 20 private patient rooms. Because of the prior success of positive deviance methodology in both units,12 positive deviance has been institutionalized in the SDU daily routine as an intervention for HH compliance since 2009. From April 1 to June 15, 2013, baseline rates of HH episodes and HAIs were established in both units. A 4-week “washout period” from June 15 to July 15 was observed for installation of the device for monitoring HH compliance using wireless technology inside the intervention unit rooms and for explaining the feedback technology to unit employees. Next, a 7-month trial of the feedback intervention was performed in the intervention unit while the other SDU served as a control. All HCWs from the intervention unit were enrolled from all shifts, and consent was not required. HH episodes were recorded by electronic handwash counters for alcohol gel (PURELL Hand Instant Sanitizer [GOJO Industries, Inc, Akron, OH], 62% ethyl alcohol þ 4% isopropyl alcohol 1 L bag). The alcohol gel dispenser (NXT 1-L model; GOJO Industries) records only 1 episode in any 2-second period even if more than 1 aliquot of alcohol is dispensed. Alcohol gel dispensers dispensed the same volume of product per use (approximately 1.3 mL) and are located inside the patient rooms and in the corridor. The total volume of product used in milliliters per patient-days and the alcohol gel aliquots (HH episodes) per patient-days were determined in both SDUs. Data were collected on a weekly basis. The total volume of product was collected every week keeping the empty bags in a box for the calculation. The study was approved by the facility’s Institutional Review Board. Feedback technology Real-time feedback technology was implemented in the intervention unit in the second phase of the study. This technology uses a wireless identification device (badge) for the HCW to record when a HCW performs HH with alcohol handrub using electronic dispensers inside the patient room. The identification devices use Zigbee technology (iHealthSys, Sao Carlos, Sao Paulo, Brazil) (wireless communication protocol based on IEEE 802.15.3

standards).13 A red light flashes above the patient bed when a HCW approaches the patient bed if HH has not been performed. A green light flashes if HH has been performed. Thus, the HCW is provided real-time feedback on HH compliance. Software integrated with a database allows reports to be generated showing how many HCWs entered the rooms, how many performed HH, and how many patients were provided care by individual HCWs, but we did not have these data during the time that the study was performed. We have data from the HH episodes that were recorded by electronic handwash counters for alcohol gel. Our study used an interesting technology, employing a HH system that uses a wall-mounted sensor to create a radiofrequency safety zone around a patient’s bed, which can detect the presence of badge-wearing HCWs near the bed. Unlike other systems, it is not necessary for the HCW to place their hands near a sensor to detect alcohol handrub on their hands.14 Our system is activated at the same time that the HCW presses the alcohol gel dispenser for HH. Detection of the signal identification badge by the electronic dispenser had a challenge: calibrating the correct distance to the signal identification badge without interference from the adjacent bed, wherein the dispenser is positioned on the same wall but near another bed in another closed room. The difficulty that we had in the project was initially to calibrate the electronics of the dispenser to prevent the detection of a badge in the adjacent bed, in the case of activation of the dispenser by a HCW without the badge. In this condition, because radio waves fail to pass through the walls of the bed adjacent the software for detection of the badge in the dispenser allows for an analysis of radio signal strength over a certain time interval, and detection of the tag is properly achieved without badge interference in the adjacent bed. This means the radio frequency could be detected in adjacent room and now the tag is properly adjusted without badge interference in the adjacent bed. Another problem that was identified was the condition in which a user without the badge drives the dispenser next to another user with the badge. The maximum distance for detection of the badge dispenser is around 1 m. A distance less than or equal to 1 m, with the activation of the dispenser by a user without the badge, the system will also detect the hygiene for the user with the badge. The detection signal of the badge within the limits of the physical space of the patient’s bed was one of the major challenges in developing the system. Before the beginning of this project, there were several cases of detection of buttons in the adjacent bed, causing a red light to flash without any people in the room with a badge. With the refinement of software and system calibration, that problem was solved. The system monitors in real time the signal strength of the radio and through statistical analysis determines whether the person with the badge is within the limits of the physical space of the bed. The range of the bedside sensor was delimited around 3 m to avoid an interference signal coming from the adjacent bed (in which a HCW may have a badge with a shorter distance than 3 m), and a physical barrier (metal plate) was added behind the radio faceplate bedside sensor. Therefore, the bedside sensor is able to detect the radio signal of the badge when the badge is in the field of vision in front of the bedside radio up to a distance of about 3 m. HAI surveillance was performed by trained infection preventionists using Centers for Disease Control and Prevention definitions15 in both units during the study. Invasive device utilization ratios (number of device-days/number of patient-days) were calculated for the duration of the study. Statistical analysis Statistical analyses were performed using SPSS 17.0 (SPSS Inc, Chicago, IL). Two analyses were performed: (1) a prospective

A.R. Marra et al. / American Journal of Infection Control xxx (2014) 1-4

analysis (intervention unit vs control unit) and (2) a before-andafter analysis in the intervention unit. Comparisons between units and phases of the study were performed using mixed linear regression models, accounting for the dependence of weekly measurements in the same location according to a first-order autoregressive correlation structure. All tests of statistical significance were 2-sided with a significance level set at .05. RESULTS In the baseline phase (preintervention period), there were 1,442 patient-days and 129,118 electronic handwashing counts in the intervention unit (106,870 in patient rooms and 22,248 in the corridor) (Table 1). In the control unit, there were 1,524 patientdays and 127,983 electronic handwashing counts (104,867 in patient rooms and 23,116 in the corridor). There were no differences between the 2 units in the mean number of dispensing HH episodes in rooms/patient-days and corridor/patient-days (74.5 vs 68.9, P ¼ .27 and 15.5 vs 15.2, P ¼ .81, respectively). In the second phase (intervention period), there were 1,938 patient-days and 205,227 electronic handwashing counts in the intervention unit (175,701 in the rooms and 29,526 in the corridor). In the control unit, there were 2,094 patient-days and 186,487 electronic handwashing counts (152,783 in the rooms and 33,704 in the corridor). There was a difference between the 2 units in the weekly mean number of dispensing HH episodes in rooms/patientdays (90.1 vs 73.1, respectively, P ¼ .001). There were no difference between the 2 units in the weekly mean number of dispensing HH episodes in corridor/patient-days (15.4 vs 16.1, respectively, P ¼ .48). There was no statistically significant difference in alcoholbased handrub weekly mean consumption (milliliters/patientdays) between the 2 units (103.1 vs 80.7, respectively, P ¼ .15) (Table 1). No differences were found in the utilization rates of central venous catheters, peripherally inserted central catheter lines, tracheostomy, bi-level positive airway pressure (bilevel ventilation), and contact precautions between the 2 study units in the second phase. For the analysis in the intervention unit comparing phases 1 and 2 (Table 2), there was a significant difference in the weekly mean number of dispensing HH episodes in rooms/patient-days (74.5 vs 90.1, respectively, P ¼ .01) but no difference between the 2 study phases in the weekly mean number of number of dispensing episodes in corridor/patient-days (15.5 vs 15.4, respectively, P ¼ .98). There was an increase in the alcohol-based handrub weekly mean consumption (milliliters/patient-days) between the 2 phases (68.9 vs 103.1, respectively, P ¼ .04). No differences were found in the utilization rates of peripherally inserted central catheter lines, urinary catheters, tracheostomy, and bi-level positive airway pressure (bilevel ventilation) between the 2 phases in the intervention unit. In the first phase of the study, there was 1 central line-associated bloodstream infection in the intervention unit (0.96/1,000 device-days) and 1 catheter-associated urinary tract infection (8.4/1,000 device-days) in the control unit. In the second phase, there was 1 central line-associated bloodstream infection (0.85/1,000 device-days) in the intervention unit. DISCUSSION In this study, we demonstrated that use of real-time feedback using wireless technology resulted in a significant increase in HH. The validity of our results is demonstrated not only by the lack of improvement in HH compliance in the control unit but no increase in HH in corridors of the intervention unit where feedback was not provided. Although HH is considered a simple intervention, it remains the most powerful infection control prevention method.1,3

3

Table 1 Comparison between the intervention and control units after implementation of real-time feedback Intervention unit Second phase Mean number of dispensing episodes in rooms/patient-days Mean number of dispensing episodes in the corridor /patient-days Mean number of total dispensing episodes/patient-days* ABHR volume consumed, mean (milliliters/patient-days) Device use rates (device-days/patient-days) Central venous line PICC line Urinary catheter Tracheostomy Contact precaution BIPAP

Control unit

Mean

SD

Mean

SD

P value

90.1

15.7

73.1

7.4

.001

15.4

3.1

16.1

18.7

.48

105.5

17.6

89.22

8.7

.002

103.1

51.9

80.7

29.6

.15

0.15 0.37 0.11 0.09 0.15 0.41

0.08 0.12 0.04 0.05 0.05 0.11

0.12 0.35 0.06 0.09 0.17 0.38

0.06 0.09 0.04 0.04 0.06 0.16

.22 .62 .001 .71 .32 .54

ABHR, alcohol-based handrub; BiPAP, bi-level positive airway pressure; PICC, peripherally inserted central catheter; SD, standard deviation. *Alcohol gel dispenser in rooms þ corridor.

Using observers for evaluating HH compliance is fraught with problems.16 As we have demonstrated, wireless communication can contribute not only to measuring HH compliance but can also effect improvement in compliance by providing real-time feedback. Many medical device manufacturers are already using this wireless technology to transmit information.9,17,18 This technology, in addition to identification, allows the transmission of information in both directions at high speed (even in remote monitoring systems). It has been suggested that radio frequency identification (RFID) technology may be useful in monitoring HH compliance; however, this technology is expensive and generates high maintenance costs, although cost is more reasonable when there are many tags for a few receivers. Another important consideration is that this technology is not yet widespread. The great attractiveness of RFID technology is that the badges do not require batteries. Other technologies, such as WiFi or Zigbee, require the use of batteries. However, the maintenance/replacement of the battery may prove easier to maintain and may be cheaper than the maintenance and hardware for RFID. Our batteries typically lasted 960 hours. We also believe that using tags in badges is much more practical than using tags in bracelets or in shoes.19 There are several limitations to this study. One is that HH compliance could only be assessed when the HCW performed HH using alcohol gel product. Moreover, we only monitored alcohol consumption data; we were not able to monitor chlorhexidine consumption because this product has been used for daily bathing of our SDU patients since 2009 and is used from the same dispenser as that used for HH. However, in our SDUs, more than 90% of the HH product consumed is alcohol gel (data not shown). One reason for not finding a strict correlation between electronic handwash counters and product volume measurement is that patients and their families inside the rooms also use the alcohol gel for HH. Because the patient is taught about the importance of using alcohol gel for HH to prevent infections, but not taught about the quality of hand disinfection, or the manner of using the electronic handwash device, it is possible that patients and family members and even HCWs pushed the dispenser multiple times in a short time period (Although the product will be dispensed on demand, only 1 episode of HH is recorded for every 2-second time period.) or when they push the dispenser once suboptimally resulting in a small dispensed volume.17

4

A.R. Marra et al. / American Journal of Infection Control xxx (2014) 1-4

Table 2 Comparison between phase 1 (control) and phase 2 (real-time feedback intervention) in the intervention unit Phase 1 Intervention unit

Mean

Mean number of dispensing episodes in rooms/patient-days Mean number of dispensing episodes in the corridor /patient-days Mean number of total dispensing episodes/patient-days* ABHR volume consumed, mean (milliliters/patient-days) Device use rates (device-days/patient-days) Central venous line PICC line Urinary catheter line Tracheostomy Contact precaution BIPAP

74.5

2

Acknowledgment

Mean

SD

P value

14.2

90.1

15.7

.01

15.5

2.4

15.4

3.2

.98

89.9

16.0

105.5

17.6

.03

References

68.9

13.5

103.1

51.9

.04

1. Boyce JM, Pittet D, Healthcare Infection Control Practices Advisory Committee; HICPAC/SHEA/APIC/IDSA Hand Hygiene Task Force. Guideline for hand hygiene in health-care settings: recommendations of the Healthcare Infection Control Practices Advisory Committee and the HIPAC/SHEA/APIC/IDSA Hand Hygiene Task Force. Am J Infect Control 2002;30:S1-46. 2. Bischoff WE, Reynolds TM, Sessler CN, Edmond MB, Wenzel RP. Handwashing compliance by health care workers: the impact of introducing an accessible, alcohol-based hand antiseptic. Arch Intern Med 2000;160:1017-21. 3. Gould DJ, Chudleigh JH, Moralejo D, Drey N. Interventions to improve hand hygiene compliance in patient care. Cochrane Database Syst Rev 2007;2. CD005186. 4. Scheithauer S, Kamerseder V, Petersen P, Brokmann JC, Lopez-Gonzalez LA, Mach C, et al. Improving hand hygiene compliance in the emergency department: getting to the point. BMC Infect Dis 2013;13:367. 5. Marra AR, D’Arco C, BravimBde A, Martino MD, Correa L, Silva CV, et al. Controlled trial measuring the effect of a feedback intervention on hand hygiene compliance in a step-down unit. Infect Control Hosp Epidemiol 2008;29: 730-5. 6. Aboumatar H, Ristaino P, Davis RO, Thompson CB, Maragakis L, Cosgrove S, et al. Infection prevention promotion based on the PRECEDE model: improving hand hygiene behaviors among healthcare personnel. Infect Control Hosp Epidemiol 2012;33:144-51. 7. Marra AR, Noritomi DT, Westheimer Cavalcante AJ, Sampaio Camargo TZ, Bortoleto RP, Durao Junior MS, et al. A multicenter study using positive deviance for improving hand hygiene compliance. Am J Infect Control 2013;41: 984-8. 8. Marra AR, Moura DF, Paes AT, dos Santos OF, Edmond MB. Measuring rates of hand hygiene adherence in the intensive care setting: a comparative study of direct observation, product usage, and electronic counting devices. Infect Control Hosp Epidemiol 2010;31:796-801. 9. Boyce JM. Measuring healthcare worker hand hygiene activity: current practices and emerging technologies. Infect Control Hosp Epidemiol 2011;32: 1016-28. 10. Boyce JM, Cooper T, Dolan MJ. Evaluation of an electronic device for real-time measurement of alcohol-based hand rub use. Infect Control Hosp Epidemiol 2009;30:1090-5. 11. Goetz T. The mental machine. Cover: The feedback loop by Thomas Goetz. San Francisco [CA]: Conde Nast, Wired; 2011:126. 12. Marra AR, Guastelli LR, de Araujo CMP, dos Santos JLS, Filho MAO, Silva CV, et al. Positive deviance: a program for sustained improvement in hand hygiene compliance. Am J Infect Control 2011;39:1-5. 13. Polgreen PM, Hlady CS, Severson MA, Segre AM, Herman T. Method for automated monitoring of hand hygiene adherence without radiofrequency identification. Infect Control Hosp Epidemiol 2010;31:1294-7. 14. Granado-Villar D, Simmonds B. Utility of an electronic monitoring and reminder system for enhancing hand hygiene practices in a pediatric oncology unit: 21st Scientific Meeting of the Society for Healthcare Epidemiology of America, Dallas, TX, 2011, Abstract 63. 15. Horan TC, Andrus M, Dudeck MA. CDC/NHSN surveillance definition of health care-associated infection and criteria for specific types of infections in the acute care setting. Am J Infect Control 2008;36:309-32. 16. Sax H, Allegranzi B, Uçkay I, Larson E, Boyce J, Pittet D. “My five moments for hand hygiene”: a user-centered design approach to understand, train, monitor and report hand hygiene. J Hosp Infect 2007;67:9-21. 17. Marra AR, Edmond MB. New technologies to monitor healthcare worker hand hygiene. Clin Microbiol Infect 2014;20:29-33. 18. Fisher DA, Seetoh T, Oh May-Lin H, Viswanathan S, Toh Y, Yin WC, et al. Automated measures of hand hygiene compliance among healthcare workers using ultrasound: validation and a randomized controlled trial. Infect Control Hosp Epidemiol 2013;34:919-28. 19. Boudjema S, Dufour JC, Aladro AS, Desquerres I, Brouqui P. MediHandTrace: a tool for measuring and understanding hand hygiene adherence. Clin Microbiol Infect 2014;20:22-8. 20. Morgan DJ, Pineles L, Shardell M, Young A, Ellingson K, Jernigan JA, et al. Automated hand hygiene count devices may better measure compliance than human observation. Am J Infect Control 2012;40:955-9. 21. Marra AR, Edmond MB. Hand hygiene: state-of-the-art review with emphasis on new technologies and mechanisms of surveillance. Curr Infect Dis Rep 2012; 14:585-91.

0.23 0.37 0.14 0.09 0.19 0.41

SD

In conclusion, we demonstrated improvement in alcohol gel usage via implementation of real-time feedback in a step-down unit. Feasibility of the technology has been demonstrated, but at present the major barrier to expansion of utilization is cost.

0.10 0.09 0.05 0.03 0.06 0.08

0.15 0.37 0.11 0.09 0.15 0.41

0.08 0.12 0.04 0.05 0.05 0.11

.04 .94 .11 .62 .07 .96

ABHR, alcohol-based handrub; BiPAP, bi-level positive airway pressure; PICC, peripherally inserted central catheter; SD, standard deviation.

In addition, our HH device is not able to discriminate among the WHO Five Moments for Hand Hygiene. We do not yet have data about the number of HCW room entries/compliance and whether this was ever validated by human observation. It was not our intention, but our HH device can discriminate who is using the dispensers (if wearing the badge) and the time when the alcohol gel was dispensed. The nurse coordinators from the SDU controlled the HCWs wearing their badges every day, but we do not have the compliance data of HCWs wearing their badges. However, we know, using the feedback loop system, that an increase in HH was only found in the patient rooms in the intervention unit and not in the alcohol gel dispenser corridors. To be clear, the idea was not to hide the feedback loop. It was not a blinded study. HCWs in the intervention unit know that they need to wear the badge and they need to know about the red/green light for making the decision for using alcohol gel dispenser. We do not have data about sustainability for using the feedback loop (Zigbee system), but we are collecting data for having it in the future. We acknowledge that HH is only 1 component of an infection prevention program and that other components are important. Although we have data on bundle implementation for deviceassociated infections in both SDUs, we know that the device utilization rate (for central venous catheters, urinary catheters, and tracheostomies) was very similar in both study phases. Our inability to determine the effect of improved HH on HAIs was impacted by the short duration of the study and the infrequent occurrence of these infections. An important consideration for expanding the use of this technology is cost.9,15,20,21 For our 20-bed SDU, the costs were approximately $50,000 (US) for the entire cost of the study, including developing the software and developing the system in the SDU. It is difficult to estimate the cost of HH monitoring because it will depend on the number of hospital beds, but we believe that, if there was widespread use of this HH technology, it would be available for a low cost. To avoiding misunderstanding, we must point out that there are many new technologies using RFID. Our system that applied the feedback loop method used the Zigbee system. The Zigbee system is a good option for places that do not have WiFi technology, and this can be one of the reasons to decrease the costs of this new technology.

The authors thank GOJO Latin America for the support to perform this work and i-HealthSys for implementing and monitoring the hand hygiene feedback loop system in the intervention unit.

The use of real-time feedback via wireless technology to improve hand hygiene compliance.

Hand hygiene (HH) is widely regarded as the most effective preventive measure for health care-associated infection. However, there is little robust ev...
202KB Sizes 0 Downloads 3 Views